INFECTION AND IMMUNITY,0019-9567/01/$04.0010 DOI: 10.1128/IAI.69.4.2692–2699.2001

Apr. 2001, p. 2692–2699 Vol. 69, No. 4

Copyright © 2001, American Society for Microbiology. All Rights Reserved.

Antimycobacterial Agent Based on mRNA Encoding Human b-Defensin 2 Enables Primary Macrophages To Restrict

Growth of Mycobacterium tuberculosisKEVIN O. KISICH,1* LEONID HEIFETS,2 MICHAEL HIGGINS,2 AND GILL DIAMOND3

Departments of Immunology1 and Medicine,2 National Jewish Medical and Research Center, Denver, Colorado 80206,and Department of Anatomy, Cell Biology, and Injury Sciences, UMDNJ-New Jersey

Medical School, Newark, New Jersey 071033

Received 15 September 2000/Returned for modification 11 October 2000/Accepted 26 December 2000

Human macrophages are hosts for Mycobacterium tuberculosis, the causative agent of tuberculosis, whichkilled approximately 1.87 million people in 1997. Human alveolar macrophages do not express a- or b-de-fensins, broad-spectrum antimicrobial peptides which are expressed in macrophages from other species moreresistant to infection with M. tuberculosis. It has been previously reported that M. tuberculosis is susceptible tokilling by defensins, which may explain the difference in resistance. Defensin peptides have been suggested asa possible therapeutic strategy for a variety of infectious diseases, but development has been hampered bydifficulties in their large-scale production. Here we report the cellular synthesis of human b-defensin 2 via high-ly efficient mRNA transfection of human macrophages. This enabled mycobactericidal and mycobacteristaticactivity by the macrophages. Although human macrophages are difficult to transfect with plasmid vectors,these studies illustrate that primary macrophages are permissive for mRNA transfection, which enabled ex-pression of a potentially therapeutic protein.

Tuberculosis continues to infect and kill approximately 2million people each year world wide. It is estimated that oneout of three humans is infected, leading to 8,000,000 new casesof active tuberculosis each year (10). The number of cases oftuberculosis is expected to double by the year 2020. Greaterknowledge of the mechanisms of resistance to this pathogen, aswell as new therapeutics, is needed. One of the first cell typesto encounter Mycobacterium tuberculosis after inhalation is themacrophage. However, M. tuberculosis multiplies rapidly incultured human macrophages even when they are stimulatedwith cytokines (9). Therefore, other elements of the immunesystem may assist macrophages in limiting the multiplication oftubercle bacilli in the approximately one-third of the earth’shuman population that is infected with M. tuberculosis but doesnot develop active disease (10).

One important element of our innate immune defensesagainst microorganisms are small antimicrobial peptides knownas defensins (6). These small (30 to 50 amino acids) cationicpeptides are found in a variety of mammalian myeloid andepithelial cells and are bactericidal or bacteristatic for a broadspectrum of microbes, including M. tuberculosis (21, 23). De-fensins are primarily divided into two subclasses, a- and b-de-fensins, based on their structural characteristics and are foundin a variety of tissues and cell types. They are among the mostabundant components in phagocytic cells, where they partici-pate in the oxygen-independent killing of ingested microorgan-isms. In epithelial cells, such as in the small-intestinal crypts(25), the female reproductive tract (27), and the trachea (8),they have been predicted to provide a first line of host defense

by acting in the luminal contents as a component of the innateimmune response. In the mammalian airway, b-defensins havebeen found in tracheal mucosa (8), nasal secretions (3) andbronchoalveolar lavage fluid (30) at concentrations which areantimicrobial in vitro, suggesting that they can perform thisfunction in vivo.

While defensins are found in rabbit (26) and bovine (28)macrophages, they are absent from human macrophages (K.O.K.and G.D., unpublished data). Although defensins have beenproposed for use as therapeutics (12), the chemical synthesis ofthese peptides is a challenge due to the complex pattern ofdisulfide bonds which stabilize their structure (18), and recom-binant methods do not produce sufficient yields (14, 31). Analternative to using defensin proteins as antimicrobial agents,using DNA to encode the defensins for intracellular expressionin a macrophage cell line, resulting in greater resistance toHistoplasma capsulatum, has been described elsewhere (4). Todate, however, there are very few reports of primary humanmacrophage transfection with DNA plasmids. Studies whichquantitate transfection efficiency report that only about 2% ofthe cells express the reporter gene (i.e., enhanced green fluo-rescent protein [eGFP]) (29, 32, 33).

We have previously observed that primary murine macro-phages efficiently accumulate RNA delivered as a complexwith cationic lipids both in vivo and in vitro and that the RNAis biologically active for at least 24 h following uptake (17). Inaddition, the RNA preferentially distributes in vivo to macro-phage-rich tissues, including the lung, which is the primary siteof infection in tuberculosis (17, 20). Therefore, we sought todetermine whether human monocyte-derived macrophages(MDM) could be efficiently transfected with an mRNA encod-ing human b-defensin 2 (HBD-2) and whether the levels pro-duced would be sufficient to inhibit the intracellular growth ofa virulent strain of M. tuberculosis.

* Corresponding author. Mailing address: Department of Immunol-ogy, National Jewish Medical and Research Center, 1400 Jackson St.,Denver, CO 80206. Phone: (303) 398-1628. Fax: (303) 398-1225.E-mail: kisichk@njc.org.

2692

on March 25, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

MATERIALS AND METHODS

Macrophage isolation and culture. Monocytes were isolated from humanwhole blood by centrifugation through Ficoll-Hypaque. The mononuclear celllayer was washed in RPMI 1640 and saline, and the monocytes were counted.Approximately 106 monocytes were dispensed into the wells of 24-well plates(Falcon; Becton Dickinson, San Jose, Calif.) and allowed to adhere for 1 h. Themonolayers were then washed three times to remove nonadherent cells. Theresulting cell monolayers consisted of ,95% monocytes as determined by hy-drolysis of the nonspecific esterase substrate fluorescein diacetate and epifluo-rescence microscopy. The few remaining nonmonocytes appeared to be lympho-cytes based on morphology. Monocyte monolayers were then cultured at 37°C for8 days to allow for differentiation into macrophage-like cells prior to infectionwith M. tuberculosis Erdman. Alternatively, cells were placed at 100,000 cells/wellinto eight-well chambered coverslips (Nalge-Nunc International, Naperville,Ill.) and allowed to adhere for 2 h in RPMI 1640 including penicillin (0.05 U/ml),streptomycin (0.05 mg/ml), L-glutamine, and 10% autologous human serum.Nonadherent cells were then removed with three washes with warm phosphate-buffered saline (PBS), and the medium was replaced with antibiotic-free Mac-rophage-SFM (Gibco-BRL, Gaithersburg, Md.). The monocytes were then al-lowed to differentiate into macrophages for 6 to 7 days at 37°C in 5% CO2.

Bacterial inoculum. To prepare mycobacterial suspensions, we collected themycobacterial lawn from the surface of Middlebrook 7H11 agar plates whengrowth had reached mid-log phase. Mycobacteria were placed into 5 ml ofMacrophage-SFM in 16-by-125-mm round-bottom borosilicate glass screw-capculture tubes with glass beads (8 to 10 mm3; Fisher Scientific) and then vortexedin pulses six times. Clumps of mycobacteria were allowed to settle at unit gravityfor 45 min. Supernatant containing a predominantly single cell suspension wasthen transferred to a new tube and allowed to settle for an additional 30 min. Thesupernatant was then transferred to 16-by-125-mm flat-bottom borosilicate glassscrew-cap culture tubes (Fisher Scientific), and the numbers of bacterial cellswere determined spectrophotometrically in a nephrometer (Becton DickinsonCrystalScan). Mycobacterial suspensions were diluted to an optical density of 1McFarland unit/ml (108 cells/ml).

Infection of macrophages. The growth kinetics of M. tuberculosis can be re-producibly measured in monolayers of human MDM when they are infected witha low inoculum in tissue culture. These procedures were performed under bio-safety level 3 conditions in the Mycobacteriology Laboratory at National JewishMedical and Research Center, Denver, Colo. This laboratory has developed andstandardized an in vitro system for testing antimycobacterial drugs (22). Thisstandardized procedure has been further developed to be used for the study ofagents which may modulate macrophage activity (K.O.K. and M.H., unpublisheddata). Macrophage monolayers were infected by replacing the medium withMacrophage-SFM containing the appropriate numbers of M. tuberculosis bacilli.Infection was allowed to continue for 1 h, after which the monolayers werevigorously washed twice with RPMI 1640-saline and incubated further in Mac-rophage-SFM.

Production of mRNA. A DNA fragment encoding eGFP was amplified fromthe retroviral plasmid pMXI-eGFP (provided by Gary Nolan, Cleveland Clinic)using PCR primers which incorporated the XbaI (59) and SacI (39) restrictionsites. The PCR product was digested with these two enzymes to mature the endsand then cloned into the SacI to XbaI sites of pSP64-poly(A) (Promega, Madi-son, Wis.). After amplification and purification from Escherichia coli, the pSP64-eGFP-poly(A) plasmid was linearized at the end of the poly(A) addition tractusing EcoRI. Capped mRNA encoding eGFP was made by in vitro transcriptionusing the Message Machine kit (Ambion, Austin, Tex.) according to a protocolsupplied by the manufacturer. The DNA template was removed by treatment ofthe reactions with DNase I for 30 min. The mRNA was purified by extractionwith phenol-chloroform-isoamyl alcohol (Ambion, Austin, Tex.), followed by theremoval of low-molecular-weight constituents by column chromatography overSephadex G-50 spin columns (NICKspin columns; Pharmacia, Uppsala, Swe-den). The resulting mRNA had an A260/A280 ratio of approximately 1.95. ThemRNA was stored at 270°C until use. HBD-2 cDNA was produced by reversetranscription-PCR using human tracheal epithelial cell mRNA as a template andpublished primer sequences (13). The cDNA was cloned into the SmaI site ofpBluescript. Templates for the in vitro transcription of HBD-2 mRNA weremade via PCR from the HBD-2 cDNA using an upstream oligonucleotide bear-ing a promoter for bacteriophage T7 RNA polymerase and a downstream oli-gonucleotide bearing a 25-residue oligo(dT) extension for the templated additionof a poly(A) tail to the in vitro transcript. In vitro transcription was carried outas described for the eGFP template.

Transfection. For the transfection of one well containing approximately100,000 cells, 2 mg of eGFP mRNA was combined with 1 mg of Oligofectin G

(Sequitur, Natik, Mass.) in 0.2 ml of serum-free, antibiotic-free RPMI 1640 in a5-ml polystyrene culture tube (Falcon). The mixture was vortexed at high speedfor 30 s and then allowed to stand at 25°C (room temperature) for 15 min. Themacrophage monolayers were washed once with PBS, and the medium wasreplaced with the mRNA-Oligofectin G complex or the yeast tRNA-OligofectinG complex (for controls) in RPMI 1640. The cultures were then returned to theincubator for 2 h, after which fetal bovine serum was added to 10%. The cellswere incubated for an additional 4 h and then fixed with neutral bufferedformalin for 30 min at 4°C. After fixation, the cells were washed extensively with1 M glycine (pH 7.2) in order to inactivate the residual formaldehyde and toretard the development of autofluorescence. The fixed cells were then allowed tostand overnight at 4°C in the dark to allow full oxidation of the eGFP chro-mophore, which is essential for the development of fluorescent properties. Thecells were examined and recorded using a Nikon Diaphot inverted microscopefitted with epifluorescence illumination and a charge-coupled device camerasystem (Nu200; Photometrics, Tucson, Ariz.). The Fluorescence intensity wasrecorded during 0.3-s exposures with a gain setting of 4 using IP Lab Spectrumsoftware (Scanalytics, Vienna, Va.). The intensity was integrated within theregion defined by the cell, and the average background of an area devoid of cellswas subtracted.

Immunohistochemistry. Cells were grown on eight-chamber slides, fixed informalin at 4°C, and washed in 1 M glycine. Immunohistochemistry was carriedout as described earlier (34), using specific HBD-2 antibody (a gift of T. Ganz).This antiserum was raised in rabbits following conjugation to ovalbumin, usingFreund incomplete adjuvant and Hunter’s Titermax. Nonimmune serum wasused as a specificity control and visualized using the Vector ABC kit (Vector-labs). Acid-fast staining of mycobacteria was carried out using Difco TB aura-mine-O stain according to the protocol supplied by the manufacturer (BectonDickinson Microbiology Systems, Sparks, Md.).

Enumeration of mycobacterial CFU. Macrophage monolayers were infectedwith M. tuberculosis Erdman at a 10:1 ratio for 1 h. Infected cells were thentransfected with increasing concentrations of mRNA encoding eGFP or HBD-2.Mycobacterial CFU were measured 4 days after infection. Infected macrophagemonolayers were lysed at the end of the growth period with 1 ml of 0.25% sodiumdodecyl sulfate (SDS) for 10 min. Wells were then scraped, and the lysate wastransferred into plastic culture tubes. Lysate was diluted with 7H9 medium toneutralize the SDS/ spread onto Middlebrook 7H11 plates for colony growth for21 days at 37°C, and then counted.

RESULTS

We first used mRNA encoding eGFP to determine the trans-fection efficiency and the optimal mRNA/lipid ratio and con-centration. Several different ratios of RNA to cationic lipidwere tested (data not shown). The ratio which provided thebest GFP expression at 2 mg of RNA per ml was tested at high-er concentrations of RNA as well. Figure 1 shows the resultsachieved with 8 mg of eGFP mRNA per ml (300 mg/ml oflipid), where .90% of macrophages exhibited fluoresence, in-dicating successful penetration of the mRNA into the cyto-plasm of most of the macrophages. The average fluorescenceintensity of the cells increased with the concentration of mRNAapplied, up to 8 mg/ml. Increasing the mRNA concentration to16 mg/ml did not result in a further increase (Fig. 2a). Figure 2bshows that the frequency of eGFP expression exceeded levelsreported for the transfection of eGFP-encoding plasmids intomacrophages by at least 40-fold (.90% positive) compared toresults previously reported for plasmids (2% positive) (14, 18).Infection with M. tuberculosis did not reduce the transfectionefficiency of eGFP mRNA into human MDM (data notshown).

Relative toxicities of mRNAs encoding GFP versus HBD-2.Cationic lipids are known to be toxic to mammalian cells at ahigh concentration (11), as are defensins (19). We thereforesought to determine the maximum dose of GFP mRNA-Oli-gofectin G complex that could be applied to the macrophagesand whether HBD-2 mRNA had greater toxicity. Figure 3

VOL. 69, 2001 NOVEL ANTIMYCOBACTERIAL AGENT 2693

on March 25, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

shows the ability of human MDM to reduce MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide] 24 hafter treatment with increasing concentrations of either GFPmRNA or HBD-2 mRNA complexed with Oligofectin G. Thereduction of MTT was not affected by the eGFP mRNA-

cationic lipid complex until the concentration exceeded 32 mg/ml. In contrast, complexes of HBD-2 mRNA and OligofectinG became toxic at 8 mg/ml, indicating that the proteins pro-duced by translation of the mRNAs differed in toxicity aspredicted.

Production of HBD-2 and association with intracellularM. tuberculosis following mRNA transfection. The ability ofHBD-2 to affect the growth of M. tuberculosis within macro-phages depends in part on the ability of the defensin protein togain access to the mycobacteria. We therefore performed im-munocytochemistry using a specific rabbit anti-human HBD-2antiserum to determine if HBD-2 protein was produced fol-lowing transfection with HBD-2 mRNA and to determinewhere in the macrophages the protein was localized. Figure 4ashows a lack of specific staining of M. tuberculosis-infected,HBD-2 mRNA-treated macrophages with preimmune rabbitserum. Figure 4b shows the presence of M. tuberculosis viaauramine-O staining and epifluorescence microscopy. Infectedmacrophages transfected with mRNA encoding eGFP andstained with anti-HBD-2 antiserum showed a similar lack ofstaining for HBD-2 (Fig. 4c and d). However, when the specificanti-HBD-2 antiserum was used with infected macrophageswhich had been treated with HBD-2 mRNA 24 h earlier,specific staining was observed within the macrophages (Fig.4e). The staining pattern is punctate and reminiscent of bacilli.A counterstain of the sample shown in Fig. 4e with aura-mine-O revealed that the cytoplasmic structures stained withanti-HBD-2 antiserum were acid-fast bacilli (Fig. 4f). Thesedata indicate that HBD-2 was produced by the macrophagesfollowing transfection with HBD-2 mRNA and that the HBD-2was able to gain access to the intracellular mycobacteria. How-ever, not all of the acid-fast bacilli were positive for HBD-2after treatment with mRNA.

Dose response of inhibition of M. tuberculosis growth. Sincethe HBD-2 produced by the macrophages was determined tobind to the intracellular mycobacteria, we next sought to de-termine if sufficient HBD-2 could be produced by the macro-phages following mRNA transfection to inhibit the growth ofM. tuberculosis. For this experiment, macrophage monolayerswere infected with M. tuberculosis Erdman at a 10:1 ratio ofbacilli to macrophages. We found this ratio to result in the in-fection of approximately 30% of the macrophages. At 1 h afterinfection, the monolayers were treated with increasing concen-trations of HBD-2 mRNA or eGFP mRNA complexed withOligofectin G ranging from 0.5 to 8 mg/ml. The monolayerswere then incubated at 37°C and 5% CO2 for 4 days, afterwhich the monolayers were lysed and the lysates were spreadon 7H11 Middlebrook plates to determine the number of my-cobacterial CFU remaining. The results are shown in Fig. 5a.The growth of M. tuberculosis was inhibited in cell monolayerstreated with 0.5 mg of HBD-2 mRNA per ml but not in cellstreated with eGFP mRNA. The growth of M. tuberculosis inthe monolayers was prevented by treatment with 2 mg or moreof HBD-2 mRNA per ml but was enhanced by treatment withthe same concentrations of eGFP mRNA. Therefore, HBD-2mRNA treatment resulted in concentration-dependent inhibi-tion of mycobacterial growth, with an MIC of approximately 2mg/ml (20 nM).

Extended duration of M. tuberculosis growth inhibition fol-lowing single administration. Following the determination

FIG. 1. Fluorescence of human-monocyte-derived macrophages af-ter mRNA transfection. (a) Phase-contrast image of MDM transfectedwith yeast tRNA as negative control for autofluorescence. Magnifica-tion, 3400. (b) Fluorescence image of panel a showing dim autofluo-rescence. (c) Fluorescence image of macrophages transfected withmRNA encoding eGFP/showing enhanced fluorescence for the major-ity of the macrophages. The images shown are representative of fiveexperiments.

2694 KISICH ET AL. INFECT. IMMUN.

on March 25, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

that macrophages transfected with HBD-2 mRNA could in-hibit the growth of M. tuberculosis, we tested the duration ofthe growth inhibition. HBD-2 mRNA was administered asdescribed above at concentrations of 2, 4, and 8 mg/ml andcomplexed with Oligofectin G. Monolayers were lysed, and thenumbers of mycobacteral CFU were determined by growth on7H11 plates 0, 2, 5, and 7 days after infection. The results areshown in Fig. 5b. Treatment with HBD-2 mRNA resulted in areduction of CFU between days 0 and 2, whereas mycobacte-rial numbers remained constant in monolayers treated witheGFP mRNA and increased in untreated monolayers. Myco-bacterium counts increased by two- to threefold between days2 and 7 in cells treated with HBD-2 mRNA but increasedapproximately 10-fold in cells treated with eGFP mRNA. My-cobacteria in untreated monolayers increased 50-fold overallbetween days 0 and 7, while numbers of mycobacteria in theHBD-2 mRNA-treated cultures did not exceed those presentat the beginning of the experiment. Therefore, the mycobac-terial growth inhibition mediated by macrophages treated witha single addition of HBD-2 mRNA lasted for at least 7 days.Upon microscopic inspection of the monolayers, cells treatedwith HBD-2 mRNA appeared much healthier, with few signsof infection at the end of 7 days, whereas untreated cells orthose which received mRNA encoding eGFP showed extensivecytopathology, with many dead cells by day 7 (data not shown).

DISCUSSION

Since the discovery of inducible defensin genes in mamma-lian cells, it has been hypothesized that their expression could

be exogenously modulated to enhance the host defense againstpathogenic microorganisms (7). In this study we demonstratethat cultured primary human macrophages can be efficientlytransfected with mRNA encoding these potentially therapeuticproteins, resulting in enhanced microbicidal activity. This

FIG. 2. Intensity and frequency of eGFP expression following mRNA transfection. (a) Fluorescence intensity of human MDM followingtransfection with increasing amounts of eGFP mRNA-Oligofectin G complex. The fluorescence intensity is represented as relative light units(RLU), with the average background of the image subtracted. (b) Frequency of eGFP-positive cells as a function of increasing concentration oftransfection complex. Each point represents the percentage of cells in three fields which are more than two standard deviations above the RNAcontrol for that concentration.

FIG. 3. Cell viability after transfection with mRNA lipid com-plexes. Cell viability was measured by incubation with 5 mg of MTT.After 24 h the absorbance of reduced MTT was measured at 585 nmfor macrophages treated with Oligofectin G–HBD-2 mRNA complex(■) or Oligofectin G-eGFP mRNA complex (}). Cell viability wasmeasured via reduction of MTT in at least three experiments, andrepresentative results are presented here.

VOL. 69, 2001 NOVEL ANTIMYCOBACTERIAL AGENT 2695

on March 25, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

FIG. 4. Association of HBD-2 with intracellular M. tuberculosis following transfection with mRNA encoding HBD-2. (a) Nomarski image(magnification, 31,000) of cells transfected with mRNA encoding HBD-2 immunostained with preimmune serum and counterstained withauramine-O to show M. tuberculosis (arrows). (b) Fluorescence image of panel a showing locations of macrophages (dim green) and M. tuberculosis(bright green). (c) Cells transfected with control mRNA and stained with anti-HBD-2 monoclonal antibody, showing a lack of specific staining forHBD-2 corresponding to the M. tuberculosis shown in panel d. (d) Fluorescence image of panel c showing locations of M. tuberculosis. (e) Nomarskiimage of cells transfected with mRNA encoding HBD-2 immunostained with specific anti-HBD-2 serum and counterstained with auramine-O. Thebrown precipitate of the diaminobenzidine appears black under Nomarski optics and partially quenches the fluorescence from auramine-O seenin panel f. (f) Fluorescence image of panel c, showing the locations of M. tuberculosis.

2696 KISICH ET AL. INFECT. IMMUN.

on March 25, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

strongly supports the hypothesis that defensins act as hostdefense molecules both in vivo and in vitro.

The efficiency of transfection observed following delivery ofan eGFP mRNA-Oligofectin G complex (.90%) was approx-imately 40-fold greater than was previously been reported forcultured human macrophages using electroporation or lipo-plex-mediated delivery of DNA reporter vectors (29, 32, 33).We have previously observed highly efficient uptake of RNA-cationic lipid complex by murine macrophages both in vitroand in vivo (17, 20). However, DNA-cationic lipid complexesare taken up by primary murine macrophages with similarefficiency (K.O.K., unpublished data), which indicates that theuptake of nucleic acid by macrophages is not the limiting stepin expression. If DNA and RNA complexed with cationic lipidspenetrate the cytoplasm at similar levels, then the inefficiencyof DNA expression relative to RNA may be related to thetransport of the DNA to the nucleus or related to the promoteractivity, neither of which are required for translation ofmRNA.

Several cationic lipid formulations were examined in thesestudies, many of which showed efficacy in delivering exogenousmRNA to the cytoplasm. However, Oligofectin G was effectiveat a lower concentration and was less toxic to the macrophagesrelative to Lipofectamine, DOTAP, Lipofectin, or DMRIE-cholesterol (data not shown). The toxicity of the HBD-2mRNA-cationic lipid complexes was greater than for the GFPmRNA-cationic lipid complex. This may be due to the re-ported toxicity of HBD-2 for mammalian cells (19) rather thanthe lipid portion of the complex. The toxicity of the GFP-cationic lipid complex observed at concentrations of .32 mg ofRNA per ml is most likely due to the complex rather than tothe mRNA or the lipid, since the components of the complexwere either not toxic in the concentration range tested (theGFP mRNA) or were only toxic at much greater concentra-

tions (Oligofectin G). Such toxicity has been reported for othernucleic acid-cationic lipid complexes (11).

Immunostaining for HBD-2 after transfection of M. tuber-culosis-infected macrophages was mainly observed associatedwith intracellular M. tuberculosis rather than in the cytoplasmof the macrophages. Mycobacteria have been reported to re-side in phagosomes which do not normally mature to lyso-somes (5). The localization of HBD-2 to this intracellular com-partment is somewhat surprising, as it is not obvious how theHBD-2 gained access to the bacilli. In the epithelial cells whereHBD-2 is normally synthesized, it is directly secreted via thetrans-Golgi and is not stored intracellularly (6). In contrast, thea-defensins are stored in cytoplasmic granules of the polymor-phonuclear leukocytes or Paneth cells (24). It is possible thatthe HBD-2 synthesized from the transfected mRNA was se-creted from the macrophages soon after synthesis. After secre-tion, the newly synthesized HBD-2 would have to gain accessto the intracellular bacilli via the endocytic process or via directpenetration of the macrophage plasma membrane, and thenthe membrane of the mycobacterium-containing phagosome.The mycobacterium-containing phagosome has also been re-ported to exchange material with the extracellular medium viathe recycling endosome compartment (2). It is therefore pos-sible that HBD-2 secreted by the macrophages reentered thecells by endocytosis and was then transported into the myco-bacterium-containing phagosome. However, since defensinshave been shown to bind to and penetrate the plasma mem-branes of mammalian cells, direct diffusion of the newly syn-thesized HBD-2 from the trans-Golgi or extracellular mediumdirectly into the phagosomes cannot be ruled out. Elucidationof the route by which HBD-2 gains access to intracellularmycobacteria will require further studies. However, followingexposure to the mycobacterium the high affinity of defensinsfor bacterial cell membranes would tend to cause the accumu-

FIG. 5. Inhibition of mycobacterial growth in MDM monolayers. (a) Growth of M. tuberculosis following transfection with mRNA encodingHBD-2 or eGFP. The percent growth or decline of the mycobacteria was calculated by dividing the average number of colonies counted on fourreplicate plates on day 4 by the average number from the time zero lysates, with the result multiplied by 100, e.g., (day 4/day 0) 3 100. The resultsare representative of three experiments. (b) Duration of growth inhibition following mRNA transfection. Macrophage monolayers were infectedwith M. tuberculosis Erdman at a 10:1 ratio for 1 h. Infected cells were then transfected with either eGFP mRNA or HBD-2 mRNA at the indicatedconcentrations. Cultures were lysed at the days indicated, and mycobacterial CFU were measured as in panel a. Error bars indicate the standarddeviation of four samples. These results are representative of four experiments.

VOL. 69, 2001 NOVEL ANTIMYCOBACTERIAL AGENT 2697

on March 25, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

lation of defensin on the surface of the bacilli. This result wasobserved, with immunostaining of HBD-2 mainly localized tothe bacilli. We have also observed high accumulation of fluo-rescently labeled human neutrophil peptide 1 on M. avium inhuman macrophages within 5 min of addition to the medium,while similarly labeled bovine serum albumin was excludedfrom the cells (data not shown).

Exposure of intracellular mycobacteria to defensins in theextracellular medium helps to explain how alveolar macro-phages, which do not normally synthesize defensins, mightutilize defensins synthesized by nearby cells, including epithe-lia and neutrophils, to limit the multiplication of M. tubercu-losis following inhalation and phagocytosis.

The growth of intracellular mycobacteria was inhibited as aresult of transfecting mRNA encoding HBD-2 but not GFP ina dose-dependent manner. The 50% inhibitory concentration(IC50) for HBD-2 mRNA was 2 mg/ml (;20 nM), which wasapproximately fourfold less than the dose which was toxic to themacrophages. It is unclear whether the IC50 represents trans-fection of 50% of the macrophages with sufficient mRNA tocompletely inhibit growth of the bacilli or whether all of themacrophages were transfected with a similar amount of HBD-2 mRNA, which was sufficient to mediate 50% inhibition ofgrowth. The inhibition of growth was robust and remainedevident for at least 7 days of culture with a single addition ofHBD-2 mRNA.

The number of viable mycobacteria in the macrophages de-clined by approximately 50% in the first 24 h after infectionwhen the macrophages were treated with HBD-2 mRNA (Fig.5). These data imply that a true bactericidal effect could po-tentially be achieved by administering the mRNA to the cul-tures at 2-day intervals. This is consistent with other data wehave gathered (not shown) using luciferase mRNA as the re-porter for murine macrophages. The dosing schedule may lenditself to further optimization for maximum antimycobacterialactivity of HBD-2 mRNA, as may the structure and chemistryof the mRNA itself. The native mRNA encoding HBD-2 con-tains relatively long 59- and 39-untranslated regions (UTRs)predicted to have extensive secondary structure of unknownfunction but which maintain extensive homology with otherb-defensins (6). Stability and translational efficiency may beimproved by replacement of the native UTRs with those fromb-globin, which is a very stable and efficiently translatedmRNA in most cell types (17). The mRNA may also be furtherstabilized by alteration of the 29OH groups (15) and by replac-ing some of the bridging phosphate groups with phosphoro-thioate groups (16) without abolishing translational activity (1).

ACKNOWLEDGMENTS

We thank Tom Ganz for the generous gift of anti-HBD-2 antiserumand Lisa Ryan, Nancy Connell, David Riches, and Peter Henson forcritical reading of the manuscript. Oligofectin-G was generously pro-vided by Tod Woolf, Sequiteur, Natik, Mass.

G.D. was supported by grants from the NIH (HL53400) and theCystic Fibrosis Foundation.

REFERENCES

1. Aurup, H., A. Siebert, F. Benseler, D. Williams, and F. Eckstein. 1994.Translation of 29-modified mRNA in vitro and in vivo. Nucleic Acids Res.22:4963–4968.

2. Clemens, D. L., and M. A. Horwitz. 1996. The Mycobacterium tuberculosisphagosome interacts with early endosomes and is accessible to exogenously

administered transferrin. J. Exp. Med. 184:1349–1355.3. Cole, A. M., P. Dewan, and T. Ganz. 1999. Innate antimicrobial activity of

nasal secretions. Infect. Immun. 67:3267–3275.4. Couto, M. A., L. Liu, R. I. Lehrer, and T. Ganz. 1994. Inhibition of intra-

cellular Histoplasma capsulatum replication by murine macrophages thatproduce human defensin. Infect. Immun. 62:2375–2378.

5. Deretic, V., and R. A. Fratti. 1999. Mycobacterium tuberculosis phagosome.Mol. Microbiol. 31:1603–1609.

6. Diamond, G., and C. L. Bevins. 1998. Beta-defensins: endogenous antibioticsof the innate host defense response. Clin. Immunol. Immunopathol. 88:221–225.

7. Diamond, G., J. P. Russell, and C. L. Bevins. 1996. Inducible expression ofan antibiotic peptide gene in lipopolysaccharide-challenged tracheal epithe-lial cells. Proc. Natl. Acad. Sci. USA 93:5156–5160.

8. Diamond, G., M. Zasloff, H. Eck, M. Brasseur, W. L. Maloy, and C. L.Bevins. 1991. Tracheal antimicrobial peptide, a novel cysteine-rich peptidefrom mammalian tracheal mucosa: peptide isolation and cloning of a cDNA.Proc. Natl. Acad. Sci. USA 88:3952–3956.

9. Douvas, G. S., D. L. Looker, A. E. Vatter, and A. J. Crowle. 1985. Gammainterferon activates human macrophages to become tumoricidal and leish-manicidal but enhances replication of macrophage-associated mycobacteria.Infect. Immun. 50:1–8.

10. Dye, C., S. Scheele, P. Dolin, V. Pathania, and M. C. Raviglione. 1999.Consensus statement. Global burden of tuberculosis: estimated incidence,prevalence, and mortality by country. WHO Global Surveillance and Mon-itoring Project. JAMA 282:677–686.

11. Freedland, S. J., R. W. Malone, H. M. Borchers, Z. Zadourian, J. G. Malone,M. J. Bennett, M. H. Nantz, J. H. Li, P. H. Gumerlock, and K. L. Erickson.1996. Toxicity of cationic lipid-ribozyme complexes in human prostate tumorcells can mimic ribozyme activity. Biochem. Mol. Med. 59:144–153.

12. Ganz, T., and R. I. Lehrer. 1995. Defensins: pharmacology and therapeutics.66:191–205.

13. Harder, J., J. Bartels, E. Christophers, and J.-M. Schroder. 1997. A peptideantibiotic from human skin. Nature 387:861.

14. Harwig, S. L., T. Ganz, and R. I. Lehrer. 1994. Neutrophil defensins: puri-fication, characterization and antimicrobial testing. Methods Enzymol. 236:160–170.

15. Heidenreich, O., F. Benseler, A. Fahrenholz, and F. Eckstein. 1994. Highactivity and stability of hammerhead ribozymes containing 29-modifiedpyrimidine nucleosides and phosphorothioates. J. Biol. Chem. 269:2131–2138.

16. Heidenreich, O., X. Xu, P. Swiderski, J. J. Rossi, and M. Nerenberg. 1996.Correlation of activity with stability of chemically modified ribozymes innuclei suspension. Antisense Nucleic Acid Drug Dev. 6:111–118.

17. Kisich, K. O., R. W. Malone, P. A. Feldstein, and K. L. Erickson. 1999.Specific inhibition of macrophage TNF-alpha expression by in vivo ribozymetreatment. J. Immunol. 163:2008–2016.

18. Lauth, X., A. Nesin, J. P. Briand, J. P. Roussel, and C. Hetru. 1998. Isolation,characterization and chemical synthesis of a new insect defensin from Chi-ronomus plumosus (Diptera). Insect Biochem. Mol. Biol. 28:1059–1066.

19. Lichtenstein, A. K., T. Ganz, M. E. Sestedt, and R. I. Lehrer. 1986. In vitrotumor cell cytolysis mediated by peptide defensins of human and rabbitgranulocytes. Blood 68:1407–1410.

20. Malone, R. W., P. L. Felgner, and I. Verma. 1989. Cationic liposome medi-ated RNA transfection. Proc. Natl. Acad. Sci. USA 86:6077–6081.

21. Miyakawa, Y., P. Ratnakar, A. Gururuaj Rao, M. L. Costello, O. Mathieu-Costello, R. I. Lehrer, and A. Catanzaro. 1996. In vitro activity of theantimicrobial peptides human and rabbit defensins and porcine leukocyteprotegrin against Mycobacterium tuberculosis. Infect. Immun. 64:926–932.

22. Mor, N., B. Simon, and L. Heifets. 1996. Bacteriostatic and bactericidalactivities of benzoxazinorifamycin KRM-1648 against Mycobacterium tuber-culosis and Mycobacterium avium in human macrophages. Antimicrob.Agents Chemother. 40:1482–1485.

23. Ogata, K. B., A. Linzer, R. I. Zuberi, T. Ganz, R. I. Lehrer, and A. Catan-zaro. 1992. Activity of defensins from human neutrophilic granulocytesagainst Mycobacterium avium-Mycobacterium intracellulare. Infect. Immun.60:4720–4725.

24. Ouellette, A. J. 1999. IV. Paneth cell antimicrobial peptides and the biologyof the mucosal barrier. Am. J. Physiol. 277:G257–G261.

25. Ouellette, A. J., and M. E. Selsted. 1996. Paneth cell defensins: endogenouspeptide components of intestinal host defense. FASEB J. 10:1280–1289.

26. Patterson-Delafield, J., D. Szklarek, R. J. Martinez, and R. I. Lehrer. 1981.Microbicidal cationic proteins of rabbit alveolar macrophages: amino acidcomposition and functional attributes. Infect. Immun. 31:723–731.

27. Quayle, A. J., E. Martin Porter, A. A. Nussbaum, Y.-M. Wang, C. Brabec,and S. C. Mok. 1998. Gene expression, immunolocalization and secretion ofhuman defensin-5 in human female genital tract. Am. J. Pathol. 152:1247–1258.

28. Ryan, L. K., J. Rhodes, M. Bhat, and G. Diamond. 1998. Expression ofbeta-defensin genes in bovine alveolar macrophages. Infect. Immun. 66:878–881.

29. Simoes, S., V. Slepushkin, E. Pretzer, P. Dazin, R. Gaspar, M. C. Pedroso de

2698 KISICH ET AL. INFECT. IMMUN.

on March 25, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

Lima, and N. Duzgunes. 1999. Transfection of human macrophages by lipo-plexes via the combined use of transferrin and pH-sensitive peptides. J. Leu-koc. Biol. 65:270–279.

30. Travis, S. M., B. A. Conway, J. Zabner, J. J. Smith, N. N. Anderson, P. K.Singh, E. P. Greenberg, and M. J. Welsh. 1999. Activity of abundant anti-microbials of the human airway. Am. J. Respir. Cell Mol. Biol. 20:872–879.

31. Valore, E. V., and T. Ganz. 1997. Laboratory production of antimicrobialpeptides in native conformation. Methods Mol. Biol. 78:115–131.

32. Van Tendeloo, V. F., H. W. Snoeck, F. Lardon, G. L. Vanham, G. Nijs,M. Lenjou, L. Hendriks, C. Van Broeckhoven, A. Moulijn, I. Rodrigus, P.Verdonk, D. R. Van Bockstaele, and Z. N. Berneman. 1998. Nonviral trans-

fection of distinct types of human dendritic cells: high-efficiency gene trans-fer by electroporation into hematopoietic progenitor- but not monocyte-derived dendritic cells. Gene Ther. 5:700–707.

33. Weir, J. P., and M. S. Meltzer. 1993. Transfection of human immunodefi-ciency virus type 1 proviral DNA into primary human monocytes. Cell.Immunol. 148:157–165.

34. Yount, N. Y., J. Yuan, A. Tarver, T. Castro, G. Diamond, P. A. Tran, J. N.Levy, C. McCullough, J. S. Cullor, C. L. Bevins, and M. E. Selsted. 1999.Cloning and expression of bovine neutrophil beta-defensins. Biosyntheticprofile during neutrophilic maturation and localization of mature peptide tonovel cytoplasmic dense granules. J. Biol. Chem. 274:26249–26258.

Editor: T. R. Kozel

VOL. 69, 2001 NOVEL ANTIMYCOBACTERIAL AGENT 2699

on March 25, 2021 by guest

http://iai.asm.org/

Dow

nloaded from